Production Database (PDB) API

The VFB Production Database (PDB) is the main Neo4j graph database containing integrated neuroanatomical data from multiple sources, including ontology data, expression patterns, image annotations, and connectivity information. It serves as the primary data store for comprehensive queries across VFB’s knowledge graph.

Browser Access

You can explore the PDB interactively using the Neo4j browser:

https://pdb.virtualflybrain.org/browser/

Data Structure Overview

The PDB contains over 1.1 million nodes and 43 million relationships, representing a comprehensive knowledge graph of Drosophila neuroanatomy. Below is an overview of the available node labels, relationship types, and property keys.

Node Labels (Gross Classifications)

The PDB uses hierarchical node labels to classify different types of entities:

Core Ontology Nodes:

• Class, Individual - OWL ontology elements
• Property - Relationships and attributes

Anatomical Classifications:

• Anatomy, Neuron, Glial_cell, Muscle, Sense_organ
• Nervous_system, Synaptic_neuropil, Ganglion, Neuromere
• Neuron_projection_bundle, Template

Cell Types and Lineages:

• Motor_neuron, Sensory_neuron, Neuroblast
• lineage_* (200+ lineage classifications for neural development)

Molecular and Functional:

• Gene, Expression_pattern, Clone, Feature
• Receptor, Channel, Transporter, Enzyme
• Neurotransmitter_receptor (Cholinergic, GABAergic, Glutamatergic, etc.)

Data and Metadata:

• DataSet, Sample, License, Person
• pub (Publications), Site, Template
• VFB, Deprecated, Individual

Dataset-Specific:

• FAFB, FlyCircuit, L1EM, FANC (Different connectome datasets)
• has_image, has_neuron_connectivity, has_region_connectivity

Relationship Types

The PDB defines 200+ relationship types for connecting entities:

Core Ontology Relationships:

• INSTANCEOF, SUBCLASSOF - Class hierarchies
• has_part, part_of - Anatomical composition
• located_in, overlaps - Spatial relationships

Connectivity Relationships:

• synapsed_to - Basic synaptic connections
• has_presynaptic_terminals_in, has_postsynaptic_terminal_in - Detailed synapse locations
• sends_synaptic_output_to_region, receives_synaptic_input_in_region - Functional connectivity
• synapsed_via_type_*_bouton_to - Specific bouton types

Developmental and Temporal:

• develops_from, develops_into - Developmental progression
• happens_during, existence_starts_during - Temporal relationships
• has_member - Group memberships

Functional and Molecular:

• expresses - Gene expression relationships
• has_role, capable_of - Functional classifications
• regulates, positively_regulates, negatively_regulates - Regulatory interactions

Metadata and Attribution:

• has_reference, created_by, has_license - Provenance
• database_cross_reference - External identifiers
• depicts - Image relationships

Property Keys

Properties provide additional metadata and attributes for nodes and relationships:

Identifiers and Labels:

• id, short_form, label, symbol, name
• iri, curie, unique_id

Descriptions and Definitions:

• definition, description, comment
• example_of_usage, elucidation

References and Attribution:

• PMID, PMCID, DOI, FlyBase
• created_by, creation_date, creator
• has_reference, miniref, pub

Spatial and Quantitative:

• center, orientation, volume, voxel
• Tbars, cell_count, cubic_micrometer_overlap
• NBLAST_score, neuronbridge_score

Quality and Status:

• confidence_value, curation_status
• is_obsolete, deprecated, hide_in_terminfo
• has_curation_status, obsolescence_reason

Technical Metadata:

• filename, link_base, thumbnail, nrrd, swc
• dataset_link, license_url, homepage

Query Examples

Note: These are direct Cypher queries for programmatic access to the PDB. The VFB web application typically uses higher-level query chains that combine multiple data sources and reasoning services. For user-friendly querying, use the VFB web interface.

Find all expression patterns for a gene

MATCH (ep:Expression_pattern)-[:expresses]->(gene:Gene {symbol: 'gene_name'})
RETURN ep

Find neurons with presynaptic terminals in a brain region

MATCH (neuron:Neuron)-[:has_presynaptic_terminals_in]->(region:Synaptic_neuropil {label: 'antennal lobe'})
RETURN neuron, region

Find neurons with specific neurotransmitter types

MATCH (neuron:Neuron)-[:INSTANCEOF]->(type:Class)
WHERE type.label CONTAINS 'GABAergic'
RETURN neuron, type

Find neurons in a specific lineage

MATCH (neuron:Neuron)-[:INSTANCEOF]->(type:Class)-[:has_member]->(lineage)
WHERE lineage:lineage_DL1
RETURN neuron, type, lineage

Find synaptic connections between neurons

MATCH (n1:Neuron)-[syn:synapsed_to]->(n2:Neuron)
WHERE n1:FAFB AND n2:FAFB
RETURN n1, syn, n2
LIMIT 10

Find neurons by morphology similarity (NBLAST)

MATCH (n1:Neuron)-[sim:has_similar_morphology_to]->(n2:Neuron)
WHERE sim.NBLAST_score[0] > 0.8
RETURN n1, n2, sim.NBLAST_score[0] AS score
ORDER BY score DESC
LIMIT 10

Find anatomical parts of a brain region

MATCH (part:Anatomy)-[:part_of]->(region:Anatomy {label: 'adult brain'})
RETURN part

Find genes expressed in a specific cell type

MATCH (gene:Gene)<-[:expresses]-(ep:Expression_pattern)-[:overlaps]->(cell:Neuron)
WHERE cell.label CONTAINS ' Kenyon cell '
RETURN DISTINCT gene, ep

Schema Details

OLS OWL Representation

The core schema follows OLS conventions for representing OWL in Neo4j:

• Classes are represented as nodes with label Class
• Individuals as nodes with label Individual
• Properties as nodes with label Property
• Relationships use Related edges with property labels

Denormalizations

For query performance, the schema includes denormalized named relationships alongside the generic Related edges. This allows efficient pattern matching without attribute filtering.

Expression Data Model

Expression patterns are modeled as:

(ep:Class:Expression_pattern)-[:SUBCLASSOF]->(:Class {label: 'expression pattern'})
(ep)-[:expresses]->(gene:Class)
(ep)<-[:INSTANCEOF]-(individual:Individual)
(ep)-[:overlaps]->(anatomy:Class)

Connectivity Data

Synaptic connections are represented as:

(neuron1:Individual)-[:synapsed_to {count: N}]->(neuron2:Individual)

Or with detailed synapse partonomy using GO terms.

Data Sources

The PDB integrates data from:

• FlyBase (ontology terms, features, annotations)
• VFB Knowledge Base (connectivity data)
• External ontologies (FBbt, GO, etc.)
• Image datasets and annotations
• Third-party connectivity datasets
• Expression pattern data

Updates and Maintenance

The PDB is updated regularly to incorporate new data releases and maintain consistency with upstream sources.